This is a repository copy of Pea protein microgel particles as Pickering stabilisers of oil-in-water emulsions: Responsiveness to pH and ionic strength.

White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/154927/

Version: Accepted Version

Article:

Zhang, S, Holmes, M, Ettelaie, R et al. (1 more author) (2020) Pea protein microgel particles as Pickering stabilisers of oil-in-water emulsions: Responsiveness to pH and ionicstrength. Food Hydrocolloids, 102. ISSN 0268-005X

https://doi.org/10.1016/j.foodhyd.2019.105583

© 2019, Elsevier. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/.

eprints@whiterose.ac.ukhttps://eprints.whiterose.ac.uk/

Reuse

This article is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) licence. This licence only allows you to download this work and share it with others as long as you credit the authors, but you can’t change the article in any way or use it commercially. More information and the full terms of the licence here: https://creativecommons.org/licenses/

Takedown

If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

1

1

Pea protein microgel particles as Pickering stabilisers of 2

oil-in-water emulsions: Responsiveness to pH and ionic 3

strength 4

Shuning Zhang, Melvin Holmes, Rammile Ettelaie, Anwesha Sarkar* 5

Food Colloids and Bioprocessing Group, School of Food Science and Nutrition, University of 6

Leeds, Leeds, LS2 9JT, UK 7

8

9

10

11

12

13

14

15

*Corresponding author: 16

Dr. Anwesha Sarkar 17

Food Colloids and Bioprocessing Group, 18

School of Food Science and Nutrition, University of Leeds, Leeds LS2 9JT, UK. 19

E-mail address: A.Sarkar@leeds.ac.uk (A. Sarkar). 20

21

mailto:A.Sarkar@leeds.ac.uk

2

Abstract 22

The aim of this study was to design plant protein-based microgel particles to create Pickering 23

emulsions (20 wt% sunflower oil, 0.05-1.0 wt% protein) and investigate the role of electrostatic 24

interactions on colloidal behaviour of such emulsions. Pea protein microgels (PPM) were 25

designed using a facile top-down approach of heat-set protein gel formation followed by 26

controlled shearing. The aqueous dispersion of PPM had hydrodynamic diameters ranging 27

from 200 to 400 nm at pH 7.0 to pH 9.0 with high negative charge (-30 to -35 mV) and pI was 28

pH 5.0. With increasing ionic concentration from 1 to 250 mM NaCl, the ζ-potential of PPM 29

changed to -8 mV due to charge screening effects, in line with theoretical calculations of the 30

electrostatic potential. The Pickering emulsions with smallest droplet sizes (d43) ~25 µm 31

exhibited excellent coalescence stability and high adsorption efficiency of PPM at the oil-water 32

interface (>98%) at pH 7.0, with the latter being supported by confocal microscopy showing 33

effective adsorption of the PPMs at the droplet surface. Adjusting the pH of the emulsions to 34

pI demonstrated aggregation of adsorbed PPM at the particle-laden interface providing a higher 35

degree of adsorption as well as enhancing inter-droplet flocculation and the shear-thinning 36

character as compared to those at pH 7.0 or pH 3.0. Charge screening effects in presence of 37

100 mM NaCl resulted in PPM-PPM aggregation and enhanced viscosity of the emulsions. 38

Findings from this study on pea protein microgels would open avenues for rational designing 39

of sustainable Pickering emulsions in the future. 40

41

Keywords: Pickering emulsion; pea protein; microgel; interaction potential; plant protein; 42

particle-stabilized interface 43

44

45

3

1. Introduction 46

In recent times, gradual dietary changes towards sustainable plant-based ingredients has 47

increased and, consequently designing food emulsions using plant proteins have been the 48

preferred direction for emulsifiers to reduce food-production associated environmental 49

footprints (Rayner, et al., 2014). Furthermore, plant proteins are appreciated by consumers as 50

they are “green”, “vegan-friendly” and are considered to be less allergenic and are less 51

expensive as compared to the most commonly used dairy proteins. A range of plant proteins, 52

such as soybean, pea, chickpea, faba bean, lentil, cowpea, French bean, sweet lupin, tomato 53

seed and zein protein (Ben-Harb, et al., 2018; Jayasena, et al., 2010; Karaca, et al., 2011; 54

Kimura, et al., 2008; Pan, et al., 2015; Sarkar, et al., 2016a) have been investigated for 55

stabilizing oil-in-water (O/W) emulsions. However, many if not most, of these plant proteins 56

have limited aqueous solubility (Makri, et al., 2006; Nesterenko, et al., 2013; Sarkar, et al., 57

2016a) and are less digestible as compared to the dairy proteins (de Folter, et al., 2012; Laguna, 58

et al., 2017). This restricts easy replacement of dairy proteins in food products by plant proteins. 59

60

In comparison to the conventional protein-stabilized emulsions, the investigation of plant 61

protein-based particles to create Pickering emulsions can be particularly interesting as these 62

emulsions involve particle-stabilization of the droplets and therefore do not need perfect 63

solubilisation of these proteins in the aqueous phase. In other words, these emulsions need the 64

particles to be partially wettable by the oil and aqueous phases. In addition, Pickering 65

emulsions have gained remarkable research interest in the food colloids community in recent 66

years due to their distinctive stability against coalescence and Ostwald ripening as well as their 67

ability to alter lipid digestion kinetics of emulsions post consumption (Aveyard, et al., 2003; 68

Binks, 2002; Dickinson, 2012; Dickinson, 2013; Sarkar, et al., 2019). Hence, there is a strong 69

4

demand in food industries to find cheaper plant protein-based Pickering stabilizer alternatives 70

that can effectively stabilize emulsion droplets for longer term against coalescence. 71

72

As compared to the extensive studies on particles derived from animal-based proteins being 73

used as Pickering stabilizers, e.g. whey protein microgels (Destribats, et al., 2014; Sarkar, et 74

al., 2016b) and lactoferrin nanoparticles or nanogels (David-Birman, et al., 2013; Gal, et al., 75

2013; Meshulam, et al., 2014; Sarkar, et al., 2018), those involving particles obtained from 76

plant proteins are fairly recent and are attracting significant research attention. For example, 77

Liu and co-authors (Liu, et al., 2017; Liu, et al., 2013; Liu, et al., 2014; Liu, et al., 2016; Peng, 78

et al., 2020; Peng, et al., 2018; Zhu, et al., 2017) investigated the ability of soy protein 79

nanoparticles aggregates (SPN) (~ 100 nm, created via heat-treatment at 95 oC for 15 min) to 80

act as Pickering stabilizers. On the other hand, Chen, et al. (2014) prepared heat denatured soy 81

protein nanogel particles at various pH values (pH 2.0-7.0) and added ions (0-200 mM NaCl). 82

In another recent study, Zhu, et al. (2018) suggested the importance of the electrostatic 83

screening by ions (100-200 mM NaCl) to improve freeze-thaw stability of Pickering emulsions 84

stabilized by soy protein-based nanoparticles. Besides the commonly used soy protein-based 85

particles, peanut protein microgel particles have been recently investigated, where these 86

microgel particles were prepared via enzyme treatment and had hydrodynamic diameters 87

ranging from 200 to 300 nm and were used to stabilize high-internal-phase Pickering emulsions 88

with 87 % oil volume fractions (Jiao, et al., 2018). Water-insoluble zein-based colloidal 89

particles and kafirin nanoparticles have also been reported as possible Pickering stabilizers. For 90

example, de Folter, et al. (2012) fabricated zein-based colloidal particles with an average 91

diameter of ~ 70 nm. Xiao, et al. (2016) used an anti-solvent precipitation method to produce 92

kafirin nanoparticles with a mean diameter of 206.5 nm. Gliadin colloidal particles (GCPs) 93

5

with average diameter of ~120 nm at acidic pH that were prepared by an anti-solvent procedure 94

(Hu, et al., 2016) were suggested as Pickering stabilisers. 95

96

Due to the significant academic and industrial interests resulting from their low cost, Liang, et 97

al. (2014) created pea protein nanoparticles for the first time. In such a process, an aqueous 98

dispersion of pea protein isolate (PPI) was adjusted at pH 3.0 to produce pea protein-based 99

particles with hydrodynamic diameter of 100-200 nm. In addition, such particles generated oil-100

in-water (O/W) Pickering emulsion with 20 wt% oil volume faction. In another study, Shao, et 101

al. (2015) found that such pea protein particle-stabilized Pickering emulsion enabled controlled 102

release of lipophilic bioactive compounds during in vitro gastrointestinal digestion. Compared 103

to the emulsion stabilized with untreated PPI, the Pickering-stabilised emulsion had a sustained 104

release behaviour due to their gel-like inter-droplet network formation. Another recent study 105

by Cochereau, et al. (2019) designed pea protein microgel particles with protein concentration 106

of 1-4 wt% at pH 6-8 via slow and modest heat treatment (i.e. 20- 40 ℃). 107

108

It is thus clear from the literature that stabilizing Pickering emulsions using plant 109

protein-based particles is a relatively recent endeavour. In particular, considering the 110

strong research interests by food industries and academic community in pea protein, it 111

is surprising that relatively little attention has been devoted to designing pea protein-112

based particles for the purpose of stabilizing Pickering emulsion droplets. Although the 113

gelation properties of pea proteins (Bora, et al., 1994; Mession, et al., 2015; Shao, et al., 114

2015), pea protein-based aggregates, such as heat-treated fibrillar aggregates (Munialo 115

et al., 2014), mixed pea globulin aggregates (Mession, et al., 2017), as well as thermal 116

aggregates from mixed pea globulin and β-lactoglobulin (Chihi, et al., 2016; Chihi, et 117

al., 2018), have been widely studied, there has only been two studies that have used pea 118

6

protein particles to prepare Pickering emulsions (Liang, et al., 2014; Shao, et al., 2015). 119

Furthermore, these two studies have been performed using pea protein gelation at just 120

one particular pH (pH 3.0), thus restricting the use of such emulsions over a wider range 121

of pH and ionic strengths. To our knowledge, there has been no study that has 122

systematically characterized the colloidal properties of thermally-crosslinked pea 123

protein microgel in the aqueous phase, as well as when present at the droplet surface, 124

and in particular also investigated the role of electrostatics on the colloidal behaviour of 125

these types of emulsions. Considering the recent demand of plant-based protein 126

particles, it is necessary to characterise the ability of such Pickering emulsions at a wide 127

range of pH and ionic strengths to understand their responsiveness to environmental 128

conditions during their processing and indeed after consumption. 129

130

Hence, in this study, pea protein has been used to design pea protein microgels (PPM) 131

using a top down approach for creating a heat-set hydrogel, followed by controlled 132

shearing to investigate their potential to stabilize O/W Pickering emulsions, which has 133

not been reported in literature to date. We hypothesized that pea protein microgel 134

already adsorbed to the interface would aggregate at the droplet surface by suitably 135

adjusting the pH to acidic pH, forming a densely packed layer of particles further 136

protecting the oil droplets against coalescence. Colloidal stability of pea protein 137

microgel particles (PPM) in aqueous phase and PPM-stabilized emulsions were 138

systematically characterized as a function of pH (pH 2.0-9.0) and ionic strength (1-250 139

mM NaCl) using particle sizing (dynamic light scattering), droplet sizing (static light 140

scattering), optical microscopy, confocal laser scanning microscopy (two dimensional 141

(2D) as well as three dimensional (3D) images), apparent viscosity, adsorption 142

efficiency assessment and ζ-potential measurements. The composition of PPM at the 143

7

interface was assessed using sodium dodecyl sulphate polyacrylamide gel 144

electrophoresis (SDS-PAGE). In addition, we calculated the interaction potentials of the 145

particles as a function of pH and ionic strengths at close separation distances to identify 146

the electrostatic contribution of these particles at the droplet surface during any droplet 147

aggregation phenomena. 148

149

2. Materials and Methods 150

2.1 Materials 151

Commercial pea protein concentrate (Nutralys S85XF) (PPC) with 85% protein content was 152

kindly gifted by Roquette (Lestrem, France). Sunflower oil was purchased from local 153

supermarket and used without any further purification. Mini-PROTEAN TGX Gels, ProtoBlue 154

Safe Colloidal Coomassie G-250 stain and all sodium dodecyl sulphate polyacrylamide gel 155

electrophoresis (SDS-PAGE) chemicals were purchased from Bio-Rad Laboratories, UK. 156

Sodium azide, Nile Red and Nile Blue were purchased from Sigma Aldrich (Dorset, UK). All 157

other chemicals were of analytical grade and purchased from Sigma-Aldrich Dorset, UK) 158

unless otherwise specified. All solutions were prepared with Milli-Q water (water purified by 159

a Milli-Q apparatus, Millipore Corp., Bedford, MA, USA) with a resistivity of 18.2 MΩ cm at 160

25 ℃. 161 162

2.2 Methods 163

2.2.1 Preparation of pea protein microgel (PPM) 164

Pea protein-based Pickering particles were prepared using slight modification of the top-down 165

method developed by Sarkar, et al. (2016b). Briefly, the PPC powder at 15 wt% powder i.e. 166

12.75 wt% protein was dispersed in 20 mM phosphate buffer at pH 7.0. The aqueous dispersion 167

8

of PPC was mixed for 2 hours using magnetic stirring at 300 rpm, and then stored at 4 oC 168

overnight. The PPC dispersion was heated at 90 oC for 1 hour to allow formation of heat-set 169

gel (Figure 1). During heat treatment, the globular pea proteins were denatured and unfolded 170

(Laguna, et al., 2017). And then the denatured proteins aggregated via the disulphide 171

crosslinking forming a macroscopic protein-based hydrogel. After cooling to room temperature 172

using flowing water at 25 oC, the pea protein hydrogel was stored at 4 oC overnight. These 173

hydrogels were mixed with 20 mM phosphate buffer (1: 1 v/v) at pH 7.0 and were broken to 174

macrogel particles using a blender (HB711M, Kenwood, UK) at speed 3 for 5 minutes. After 175

removing the air bubbles generated during blending using vacuum (25-30 bar) for 15 min, the 176

macroscopic gel particle dispersion was homogenized using two passes through a two-stage 177

valve homogenizer (Panda Plus 2000, GEA Niro Soavi Homogeneizador Parma, Italy) 178

operating at first/ second stage pressures of 250/ 50 bars, respectively. The resultant particles, 179

termed as pea protein microgels (PPM) contained 6.375 wt% protein. 180

181

2.2.2 Preparation of PPM-stabilized Pickering emulsions (PPM-E) 182

Sunflower oil was mixed with appropriate quantities of PPM at 20:80 oil: protein (w/w) ratio 183

using rotor-stator (L5M-A, Silverson machines, UK) mixing at 8,000 rpm for 5 minutes. The 184

PPM dispersion was diluted using phosphate buffer to have 0.05, 0.10, 0.25, 0.50 and 1.0 wt% 185

protein content in the final emulsions, henceforth, such emulsions are referred as E0.05, E0.1, 186

E0.25, E0.5 and E1.0, respectively. The pre-homogenized oil-PPM mixture was further 187

homogenized by a two-stage valve homogenizer (Panda Plus 2000, GEA Niro Soavi 188

Homogeneizador Parma, Italy) operating at two stages, of 250 and 50 bar pressures (Figure 1) 189

resulting in Pickering emulsions (PPM-E, E0.05-E1.0). Sodium azide (0.02 wt%) was added 190

as an antimicrobial agent. 191

9

2.2.3 Coalescence stability of PPM-E during storage 192

The PPM-Es (E0.05, E0.10, E0.25, E0.50 and E1.0) were stored at 4 oC for a period of 28 days 193

and were monitored using droplet sizing, ζ-potential measurement and visual observation to 194

select the most stable emulsions for pH and ion treatment. 195

196

2.2.4 pH and ion treatment 197

To investigate the colloidal stability of PPM and PPM-E (E1.0, chosen based on the 198

coalescence stability study) under environmental conditions, the samples were subjected to 199

different pH and ionic strengths. The pH adjustment (pH 2.0-9.0) was done by both “low to 200

high” and “high to low” methods by adding 1 M HCl or 1 M NaOH, without any salt addition 201

(Adal, et al., 2017). For “high to low” pH adjustment, the PPM at pH 9.0 was rapidly adjusted 202

to a target pH while mixing. For “low to high”, the PPM at pH 2.0 was adjusted to a target pH 203

quickly while mixing. In case of E1.0, the pH responsiveness was checked at pH 3.0, 5.0 and 204

7.0. For ionic strengths, the pH value of PPM was kept constant at pH 7.0 and ionic strength 205

was adjusted from 1-250 mM NaCl. For E1.0, the physicochemical stability was assessed by 206

subjecting the emulsions to 1 mM, 10 mM and 100 mM NaCl, respectively. 207

208

2.2.5 Size and ζ-potential measurements. 209

The hydrodynamic diameters of PPM dispersion at various pH and ionic strengths were 210

measured at 25 oC using a Zetasizer Nano-ZS (Malvern Instruments Ltd, Malvern, 211

Worcestershire, UK). The PPM sample was diluted to 0.004 wt% particle concentration with 212

Milli-Q water before measurement. Assuming that there is no particle–particle interaction in 213

the diluted sample, the hydrodynamic diameter (Dh) of the droplets was calculated using the 214

Stokes–Einstein equation (Equation 1): 215

10

𝐷ℎ = 𝑘𝐵𝑇3𝜋𝜂𝐷𝑡 (1) 216 where Dt is the translational diffusion coefficient, kB is Boltzmann’s constant, T is 217

thermodynamic temperature, and η is dynamic viscosity. The refractive index of PPM was set 218

at 1.52. The absorbance of the protein was assumed to be 0.001. 219

Droplet size distribution of PPM-Es (E0.05, E0.10, E0.25, E0.50 and E1.0) as a function of 220

storage and E1.0 as a function of different pH or with various ionic strengths were determined 221

using static light scattering (SLS) techniques using a Malvern MasterSizer 3000 (Malvern 222

Instruments Ltd, Malvern, Worcestershire, UK) at 25 ℃. Samples were added dropwise in 223 distilled water until the instrument gave an obscuration rate between 4 and 6%. The average 224

droplet size of the emulsion was reported as the volume mean diameter (d43= ∑ 𝑛𝑖𝑑𝑖4∑ 𝑛𝑖𝑑𝑖3 ) and 225 surface mean (d32= ∑ 𝑛𝑖𝑑𝑖3∑ 𝑛𝑖𝑑𝑖2) 226 where, ni is the number of droplets with a diameter, di. The refractive index of sunflower oil 227

and the dispersion medium were set at 1.46 and 1.33, respectively. The absorbance value of 228

the emulsion droplets was set at 0.001. 229

230

Zetasizer Nano ZS was used to measure the ζ-potential of PPM dispersion, PPM-Es (E0.05, 231

E0.10, E0.25, E0.50 and E1.0) as a function of storage and E1.0 as a function of pH and ionic 232

strength. Before measurement, PPM dispersion at different pH values was diluted to 0.004 wt% 233

particle concentration and E1.0 with different pH values was diluted to 0.005 wt% droplet 234

concentration using Milli-Q water adjusted to relevant pH (2.0-9.0). Similarly, PPM or E1.0 235

containing different concentrations of NaCl was diluted Milli-Q water adjusted to relevant 236

NaCl concentrations (1-250 mM). The diluted sample was then added into a folded capillary 237

cell (DTS1070 cell, Malvern Instruments Ltd., Worcestershire, UK), which had two electrodes. 238

11

After 120 seconds of equilibration in the Zetasizer at 25 oC, the PPM particles or E1.0 droplets 239

moved towards oppositely charged electrodes. The magnitude of ζ-potential was determined 240

from the terminal speed of the particle motion using Henry’s equation, with Smoluchowski 241

approximation appropriate here since the thickness of the diffused double layer is expected to 242

be much smaller than the size of the particles. 243

244

2.2.6 Adsorption efficiency of PPM at the oil-water interface 245

To determine the adsorption efficiency of PPM at the oil-water interface after pH-treatment 246

(adjusting the pH of PPM-E to pH 3.0, 5.0 and 7.0) or salt-treatment (100 mM NaCl), the 247

quantity of PPM in the emulsion phase was determined (Araiza-Calahorra, et al., 2019; Sarkar, 248

et al., 2016b). Briefly, PPM-E (E1.0) at different pHs (pH 3.0, 5.0 and 7.0) and ionic strength 249

(100 mM NaCl) were diluted (1:4 w/w) with phosphate buffer (pH 7.0) or Milli-Q water 250

(adjusted to pH 5.0 or pH 3.0) or phosphate buffer at H 7.0 containing 100 mM NaCl. All 251

diluted emulsions were centrifuged for 40 mins at 10,000 rpm, 20 oC (Fresco 21 centrifuge, 252

Thermo Fisher Scientific, Germany). The subnatants were carefully collected using a syringe 253

and then measured using a DC protein assay kit (Bio-Rad Laboratories, Hercules, CA) on UV-254

Vis Spectrophotometer with an adsorption wavelength of 750 nm. The adsorption efficiency 255

of PPM at the interface was calculated by subtracting the amount of PPM in the subnatant from 256

the total amount of PPM used initially to prepare the emulsions as a percentage of total protein 257

concentration in the emulsion. 258

259

2.2.7 Apparent viscosity measurement 260

The apparent viscosity of PPM-E (E1.0) as a function of pH and ionic strength were determined 261

at 25 °C using a Kinexus ultra rheometer (Malvern Instruments Ltd, Malvern, UK). The 262

12

apparent viscosities (ηa, Pa s) were recorded as a function of shear rates ranging from 0.1 to 263

1000 s−1. All measurements were done in triplicates and were reported as the mean and standard 264

deviation. In order to determine the flow type the emulsions as a function of pH and ionic 265

strength, the flow curves were fitted using power-law model (Ostwald-de Waele model) (see 266

Equation (2)): 267

𝜂𝑎(�̇�) = 𝐾(�̇�)𝑛−1 (2) 268 where, K is consistency coefficient (Pa sn), n is power-law index and �̇� is shear rate (s-1). 269 270

2.2.8 Microscopy 271

Optical microscopy (Nikon, SMZ-2T, Japan) was used to observe the microstructure of PPM 272

and E1.0 as a function of pH and ionic strengths. The samples undergoing optical microscopy 273

needed to be diluted 10 times in respective buffer. Zeiss confocal microscope (Model LSM 274

700, Carl Zeiss MicroImaging GmbH, Jena, Germany) was used for microstructural 275

characterization of PPM at the interface of E1.0 droplets. The oil droplets in E1.0 were stained 276

with 100 μL of Nile Red (2% w/v in in dimethyl sulfoxide) and the protein stabilizing the oil 277

droplets was stained with 500 μL of Nile Blue (10% w/v in Milli-Q water). Nile Red was 278

excited by at 488 nm whereas Nile Blue was excited at 635 nm. The stained samples were 279

mixed with an appropriate amount of xanthan gum (1 wt %) to reduce the Brownian motion of 280

oil droplets. The prepared samples were placed onto a microscope slide, covered with a cover 281

slip and observed at 63 × (oil) magnifications. 282

283

2.2.9 Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) 284

Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing 285

conditions was used to determine the composition of protein in the initial pea protein 286

13

solution, PPM dispersions and the adsorbed phase of PPM-E (E1.0). The samples included 287

PPM dispersions that are pH-treated (adjusting the pH of PPM to pH 3.0, 5.0 and 7.0) and 288

salt-treated (100 mM NaCl) PPM dispersions as well as PPM-E (E1.0) at different pHs (pH 289

3.0, 5.0 and 7.0) and ionic strength (100 mM NaCl). To determine the protein compositions 290

of absorbed phase i.e. the particles at the interace, all the PPM-E samples after pH and ionic 291

treatments were centrifuged at 14,500 g for 45 min, and then the cream phases were 292

carefully collected, dispersed in Milli-Q water and centrifuged again for 45 min at 14,500 g. 293

Approximately, 65 μL of pea protein dispersion, PPM samples and cream layer of PPM-E 294

(E1.0) samples were mixed with 25 μL pf SDS loading buffer (62.5 mM Tris-HCl, pH 6.8, 295

2% SDS, 25% glycerol, 0.01% bromophenol blue) and 10 μL of diothiothreitol (DTT, of a 296

final concentration of 50 mM), heated at 95 ℃ for 5 min.The SDS-PAGE was carried out 297 by loading 5 μL of protein marker and 10 μL of these samples-SDS buffer mixtures in the 298

Mini-PROTEAN 8-10% TGX Gels in a Mini-PROTEAN II electrophoretic unit Bio-Rad 299

Laboratories, Richmond, CA, USA). The resolving gel contained 16% acrylamide and the 300

stacking gel was made up of 4% acrylamide. 301

302

The running process had two stages; one at 100 V for 10 min at first stage and then 200 V 303

of about 20 min for the second stage. After the run, the gel was stained for 2 hours using 304

Coomassie Blue solution, which consisted of 90% ProtoBlue Safe Colloidal Coomassie G-305

250 stain and 10% ethanol. The gel was then destained using Milli-Q water overnight and 306

scanned using the ChemiDoc™ XRS+ with image LabTM Software (Bio- Rad Laboratories, 307

Inc, USA). Each band within the lanes was selected automatically by the software to cover 308

the whole band. Background intensity was subtracted after scanning an empty lane. The 309

SDS-PAGE experiments were carried out in triplicates and band intensities was reported 310

as an average and standard deviation of three reported readings. 311

14

2.2.10 Statistical analysis 312

All experimental results were reported as mean and standard deviations of five measurements 313

on triplicate samples (n = 5 × 3). The statistical analyses were conducted using one-way 314

ANOVA and multiple comparison test using SPSS software (IBM, SPSS statistics, version 24) 315

and the significant difference between samples were considered when p < 0.05. 316

317

3. Results and Discussion 318

3.1. Characteristics of aqueous dispersion of PPM 319

The aqueous dispersion of PPM in phosphate buffer at pH 7.0 had a narrow size distribution 320

with single peak in the size range of 100 to 1000 nm (Figure 2). The PPM had a hydrodynamic 321

diameter (Dh) of about 232 nm (polydispersity index (PDI) < 0.2) (Figure 3a). Although an 322

aqueous dispersion of PPC showed multimodal size distribution with high PDI of nearly 1.0 323

(data not shown), preparing PPM via the top-down approach of heat-set gel formation and 324

controlled shearing appears as a feasible approach to create plant protein-based particles with 325

high colloidal stability. For instance, the PPM were highly negatively charged (-40 mV) 326

(Figure 3b) showing no particle aggregation or macroscopic sedimentation (Figure 3c) at pH 327

7.0, respectively over a year storage. 328

329

3.2 Colloidal stability of aqueous dispersions of PPM 330

As shown in Figure 3a, Dh of PPM ranged from 200 to 400 nm at neutral to alkaline pH (pH 331

7.0 to pH 9.0) (p > 0.05). However, at pH 6.0, PPM showed a marked increase in Dh supported 332

by correspondingly steep increase in PDI to 0.8 as compared to that at neutral pH (p < 0.05). 333

The Dh of PPM dispersions increased to the highest values of > 8,000 nm at pH 5.0 followed 334

15

by a decrease to < 2,000 nm at 3.0 < pH < 5.0 (Figure 3a). High degree of particle aggregation 335

and sedimentation observed using optical microscopy and macroscopic images, respectively, 336

indicate pH 5.0 to be the isoelectric point (pI) (Figure 3c). Caution should be exercised while 337

interpreting these Dh results with values above 1000 nm, when using dynamic light scattering 338

(Supplementary Figure S1). Thus, focusing on the trend of increasing values of Dh, the 339

colloidal stability of aqueous dispersions of PPM was limited at pH 6.0 with high PDI (Figure 340

3a) and excessive particle aggregation (Figure 3c). 341

342

The colloidal behaviour of PPM as a function of pH was investigated using ζ-potential 343

measurements of PPM (Figure 3b). The ζ-potential of PPM increased from -40 mV to +30 mV 344

with pH change, from pH 9.0 to pH 2.0 (p < 0.05). Noticeably, the net surface charge was close 345

to zero at pH 5.0 (Figure 3b) validating this to be the pI, corroborating with the largest 346

hydrodynamic diameter and PDI (almost close to 1.0) data (Figure 3a). At this pH, the 347

positively-charged amino-groups in PPM were balancing the negatively-charged carboxyl 348

groups. Interestingly, the pI of PPM (Figure 3b) shifted slightly from that of the PPC, where 349

latter is reported to be around 4.0 (Adal, et al., 2017). Such discrepancy in pI values between 350

protein concentrates and protein microgel particles are possible owing to the unfolding process 351

during thermal treatment of the latter and consequently exposure of some charged groups. At 352

or below pH 4.0, PPM showed net positive charge ranging from +20 to +30 mV. An interesting 353

study by Destribats, et al. (2014) demonstrated that in whey protein microgel particles (WPM), 354

some larger particle were present even when the pH value was far from the pI and the particles 355

possessed electrostatic charge. The larger particles were postulated to be associated with the 356

swelling of WPM i.e. the solvation of the exposed protein groups enabling WPM to swell. 357

Nevertheless, in the present study even if some degree of swelling of PPM might have occurred, 358

16

due to the solvation of the protein groups that was promoted at acidic pH, such effects have 359

been overshadowed by the dominant particle aggregation as observed in Figure 3c. 360

361

It is worth noting that the PPM in presence of electrolytes showed no significant difference in 362

Dh and PDI (Figure 4a) as a function of ionic strengths. This was in close agreement with no 363

aggregation or sedimentation observed in optical microscopy or macroscopic images (Figure 364

4c). However, a progressive decrease of ζ-potential magnitude from -31 to -7.5 mV was 365

observed (Figure 4b) as the ionic strength increased from 1 to 250 mM NaCl. Similar salt-366

induced reduction in ζ-potential in plant protein-based particles has been observed in the case 367

of soy protein-based nanoparticles, where Liu, et al. (2013) demonstrated that the increasing 368

NaCl concentration (0-500 mM) led to the decrease of absolute magnitude of ζ-potential of soy 369

protein nanoparticles. This is due to a decreased Debye length which for constant charged 370

surfaces at any rate will lead to a lower ζ-potential. 371

372

Colloidal particles, such as PPM in this study, dispersed in aqueous solutions will experience 373

Derjaguin-Landau-Verwey-Overbeek (DLVO) forces, other types of short-ranged, attractive 374

non-DLVO forces (Hogg, et al., 1966) as well as possible steric repulsion. To understand the 375

role of interaction in the colloidal stability of PPM, we used DLVO theory to calculate the 376

inter-particle electrostatic-mediated activation energy barrier between sub-micron sized 377

PPMs, as two such particles approach each other at varying pHs and ionic strengths (Figure 378

5). This is done using equation (3) (Cosgrove, 2010): 379

UR=2πεψ2aln[1+exp(-kh)] (3) 380

for the electrostatic component of particle-particle interaction. For equation (3), ε is the 381

permittivity of the system, i.e. ε0εr (ε0 is the permittivity of vacuum, and εr is relative 382

17

permittivity of water), ψ is the surface potential approximately equal to the ζ-potential values 383

measured at different pHs or ionic strengths, a is the radius of PPM particle (i.e. 118 nm), h is 384

the particle-particle surface separation distance ranging from 0.1 to 5 nm and κ-1 is the Debye 385

length, which was calculated from equation (4): 386

κ =[𝑁𝐴𝑒2𝜀𝑘𝑇 ∑ 𝑧𝑖2𝑐𝑖∞𝑖 ]12 (4) 387 Here, 𝑧𝑖 is the valency of ith type of ion, 𝑐𝑖∞ is the number density of that ion, NA is the 388 Avogadro’s number, and T is the temperature (298 K). 389

The total interaction potential (U) between PPMs is the summation of electrostatic repulsion 390

(UR) and van der Waals attraction (UVW). In this calculation, the pH independent van der Waals 391

attractive energy (UVW) between two equal sized PPM was calculated using equation (5): 392

UVW = − 𝑎𝐴𝐻12ℎ (5) 393 where, AH is the Hamaker constant which has been assumed to be 1 kBT, similar to other 394

reported protein particles (Tuinier, et al., 2002). 395

The total interaction potential was higher than 100 kBT at pHs < 3.0 and pHs > 6.0 (Figure 5a), 396

suggesting an electrostatically-induced energy barrier being sufficient to slow down any 397

aggregation to negligible rates (see Supplementary Figure S2a for electrostatic repulsion). 398

However, the van der Waal’s attractive forces particularly for PPM in the acidic pH i.e. pH 399

2.0-4.5 played a dominant role when at close PPM-PPM separation distance (< 0.15 nm). This 400

caused the total energy maximum to fall below ~10 kBT (Figure 5a), in line with high Dh, high 401

PDI (Figure 3a) and extensive particle aggregation (Figure 3c). The ζ-potential of PPM 402

reflected minimal surface charge (almost close to zero, Figure 3b) at pH 5.0 and the 403

correspondingly lack of sufficient electrostatic repulsion barrier, to provide a large energy in 404

18

the presence of strong van der Waal’s attraction (Figure 5a). A low energy barrier accelerated 405

the aggregation of the particles (Figures 3a and 3c). 406

407

The UR and UVW between PPMs with varying ionic strengths were also calculated using 408

equations (3-5), as shown in Figure 5b (see Supplementary Figure S2b for UR and UVW). 409

Interestingly, at 1-10 mM NaCl, the interaction was mainly dominated by electrostatic 410

repulsive forces with PPM-PPM interaction potential of > 100 kBT overshadowing effects of 411

van der Waal’s attractive forces (Figure 5b) over most separation distances. As expected, the 412

electrostatic repulsion between particles was screened, down to nearly zero, with the increase 413

in the NaCl concentration (Figure 5b). Once the salt concentration was above 50 mM, the 414

dominating van der Waal’s attractive forces between PPMs should result in particle 415

aggregation, leading to an unstable dispersion. However, this was not the case experimentally. 416

In fact, larger particle aggregates were not observed (Figure 4c), and the size of PPM was 417

stable at a range of 230 to 240 nm, with lower PDI values (Figures 4a). The possible 418

explanation for the discrepancy between the theoretically predicted aggregation and an 419

experientially observed stable dispersion might be attributed to the steric repulsion effects 420

associated with the hairy PPM particles, which is not strongly influenced by electrolyte 421

concentration. This requires experimental investigation in the future. This behaviour is unlike 422

that of WPM, soy protein nanoparticles, zein colloid particles and kafirin nanoparticles (de 423

Folter, et al., 2012; Destribats, et al., 2014; Liu, et al., 2013; Xiao, et al., 2016), where NaCl 424

addition is known to cause aggregation of particles in the aqueous dispersions. 425

426

3.3 Characteristics of Pickering O/W emulsions stabilized by PPM (PPM-E) 427

3.3.1 Stability of PPM-E prepared using various concentration of PPM 428

19

To determine the optimum concentration of PPM to stabilize PPM-E, coalescence stability of 429

PPM-Es stored under refrigerated conditions for a period of 28 days were characterized using 430

size, charge and visual imaging of any oiling off. Table 1 shows the emulsion droplet size 431

and ζ-potential, and Figure 7 presents the corresponding visual images of the freshly prepared 432

PPM-Es (E0.50-1.0) after 7, 14 and 28 days storage, respectively, at 4 ℃. The cream layer was 433

evident in all the freshly prepared emulsions (E0.05-0.5), except in E1.0 (Figure 7), however, 434

all the emulsions reverted back to a homogenous dispersion after gentle hand shaking without 435

any visual evidence for coalescence. After a week of storage, although no phase separation was 436

discernible in E0.05 and E0.1 macroscopically (Figure 7), larger oil droplets were visible after 437

diluting the emulsions with buffer (Supplementary Figure S3) with clear signs of coalescence 438

and hence the size and ζ-potential could not be measured (Table 1) in these emulsions. This is 439

expected as the insufficient quantities of PPM in E0.05 and E0.1 prevented sufficient particle 440

coverage to enable stabilization of the large surface area of such high fraction of oil droplets 441

(20 wt%). The network formed by PPM in continuous phase appears to encapsulate the oil 442

droplets, and this prevented the macroscopic oil phase separation (Figure 7). 443

444

The remaining three emulsions (E0.25, E0.5 and E1.0) did not show any coalescence upon 445

dilution within the first week of storage (Supplementary Figure S3) and the ζ-potential values 446

ranging from -38 to -41 mV (Table 1) indicated that the droplets had high net negative surface 447

charge providing sufficient electrostatic stability to the droplets. After two weeks of storage, 448

emulsions E0.25 and E0.5 showed an increase in droplet size (p < 0.05), however, no change 449 in droplet size was observed in the case of E1.0 (p > 0.05) (Table 1). This was associated with 450 decreasing values of ζ-potential for E0.25 and E0.5 (p < 0.05) and no significant change in ζ-451

potential value in the case of E1.0 (p > 0.05) over time (Table 1). More importantly, after 452 storage over a month, oil droplets were detected in the diluted E0.25 and E0.5 emulsions 453

20

(Supplementary Figure S3). There was no significant difference in droplet size and ζ-454

potential of emulsion E1.0 over this 28 day storage period (p> 0.05), which is also in line with 455 results of other Pickering emulsions, where 1 wt% of particles were needed to provide 456

sufficient surface coverage of the droplets (Araiza-Calahorra, et al., 2019; Du Le, et al., 2020). 457

In summary, 1.0 wt% PPM was sufficient to create small-sized (d43 ~25 μm) stable droplets 458

that showed excellent resilience against coalescence (E1.0). Hence, hereafter, 1 wt% PPM was 459

chosen as the optimum concentration to create Pickering emulsions (E1.0) and test their 460

response to pH and ionic strengths. 461

462

3.3.2 Composition of proteins in the adsorbed phase of E1.0 463

Compositions of PPC, PPM dispersions and the adsorbed phase of E1.0 were analysed by 464

reduced SDS-PAGE to understand if there was any difference in composition of the protein 465

subunits in bulk phase and those that were adsorbed at the interface in case of E1.0. The protein 466

mixture in PPC or the laboratory-synthesized PPM had more than ten polypeptides (Figure 6) 467

and the bands at range of 36-42 kDa had the highest proportion. The polypeptide composition 468

of PPC or PPM was similar to those reported by Mession et al. (2015) and Peng, et al. (2016) 469

containing the three main proteins, legumin, vicilin and convicilin. Convicilin was the band at 470

66 kDa. Legumin is a globular protein with an acidic subunit (Lα) at about 42 kDa and basic 471

subunit (Lβ) at 18- 24 kDa. Vicilin proteins consisted of three subunits (Vi1-3), which are 472

respectively observed in fractions of around 50 kDa, 36-30 kDa and 20 kDa. Overall, SDS-473

PAGE results indicate that the formation of PPM from PPC involved all protein subunits, which 474

is in accordance with results obtained previously that heat treatment did not affect the protein 475

composition (Laguna, et al., 2017). In addition, the adsorbed phase of E1.0 showed similar 476

molecular weight profiles to that of PPM that were not influenced by pH-treatment or exposure 477

to ions. This suggests that adsorption or environmental stresses (pH or ions) applied to PPM 478

21

had limited effect on the protein composition of the particles. In addition, a significant 479

proportion of PPM did not enter the resolving gel and were retained in the stacking gel suggest 480

that they were oligomers above 250 kDa. This suggests that DTT was not sufficient to break 481

covalent disulphide bridges in the PPM effectively. 482

483

3.3.3 Influence of pH on behaviour of E1.0 droplets 484

Figure 8a shows the mean droplet size distribution of E1.0 (20 wt% sunflower oil) as a function 485

of pH (pH 3.0, pH 5.0 and pH 7.0) with corresponding Sauter mean diameter (d32), De 486

Brouckere mean diameter (d43) and ζ-potential shown in Table 2. The initial E1.0 (pH 7.0) has 487

a significantly larger proportion of droplets in the peak area of 10-100 μm. The smaller peak 488

in the range of 0.1-1 μm as observed in Figure 8a, overlaps neatly with the size distribution of 489

PPM estimated using dynamic light scattering (Figure 2), suggesting that these small particles 490

in Figure 8a might be the free microgel particles that were not adsorbed to the droplet surface 491

during the homogenization process (Sarkar, et al., 2016b). Comparing the mean diameter of 492

the PPM particles of ~230 nm (Figure 3a) and the mean size of the oil droplets (d43) of ~ 25 493

μm (Table 2), the ratio of oil droplet size to PPM size ranges from 100:1 to 1000:1, which is a 494

signature of a classical Pickering emulsion (Ettelaie, et al., 2015; Sarkar, et al., 2016b). The ζ-495

potential of E1.0 is about -41 mV (Table 2), similar to that of PPM (Figure 3b) at pH 7.0. This 496

suggests that perhaps a monolayer of PPM might be present at the surface of droplets. 497

498

Interestingly, after adjusting the pH of E1.0 to pH 5.0 or pH 3.0, the d43 values were comparable 499

to that of E1.0 at neutral pH (p > 0.05). However, d32 values of E1.0 at different pHs showed 500

a significant difference (p < 0.05) (Table 2). This might be attributed to the fact that d32 value 501

was more affected by the changes in free (unadsorbed) microgel peak as a function of pH as 502

22

compared to d43 value, which is line with the behaviour of PPM in aqueous phase as observed 503

in Figure 3a. The width of the peak at the range 10-100 μm in the droplet size distribution was 504

narrower with a higher peak height at a lower pH (pH 3.0) than that seen for E1.0 at pH 7.0 505

(Figure 8a). Also, one might not expect such narrow droplet size distribution in E1.0, 506

particularly at pH 5.0 considering it is the isoelectric point (pI) of PPM, where E1.0 droplets 507

will also possess negligible surface charge (Table 2). This is a unique behaviour, unlike that 508

observed in PPM in the aqueous phase (Figures 3a and 3c) as well as previous studies, where 509

pH adjustment of emulsions stabilized by pea protein isolate to lower pH increased the 510

emulsion droplet size and reduced the emulsion stability (Adebiyi, et al., 2011). 511

512

To understand the aggregation behaviour, the apparent viscosities of the emulsions (E1.0) at 513

pH 7.0, pH 5.0 and pH 3.0 were determined using shear rate ranging from 0.1 to 1000 s−1 514

(Figure 8b). The Ostwald de Waele model was applied to fit the flow curves and the 515

corresponding fit parameters (consistency coefficient (K), flow index (n), regression coefficient 516

(R2) were summarized in Table 3. The R2 of all samples was 0.98, confirming a good fit to 517

the model. For E1.0 at different pH, n varied from 0.63 to 0.80, suggesting that E1.0 was a 518

pseudoplastic fluid showing shear-thinning behaviour at all the tested pH values. Emulsions at 519

pH 5.0 showed the highest K and the lowest n (p 0 .05) 525

in either K or n as compared to those at pH 7.0 (p > 0.05) (Table 3) suggesting that the droplet 526

flocs that were broken down in the direction of shear were similar at pH 7.0 and pH 3.0. 527

23

Visual images of E1.0 after 3 months of storage showed no distinct oil layers again confirming 528

the ability of PPM to to act as effective Pickering stabilizer (Figure 9a). Confocal laser 529

scanning two-dimensional (2D) (Figure 9a) and three-dimensional (3D) (Figure 9b) 530

micrographs revealed a thick layer of PPM (Nile Blue staining the protein microgels, displayed 531

in green) adsorbed at oil-water interface (Nile Red staining the oil droplets, displayed in red), 532

acting as a barrier to coalescence as observed visually. The confocal micrographs showed 533

evidence of bridging flocculation as pH was reduced to pI (pH 5.0) with visual signs of 534

creaming. This is in agreement with the higher viscosities and consequently higher consistency 535

coefficients of the emulsions at pH 5.0 as compared to those at pH 7.0 and pH 3.0 (p < 0.05) 536

(Figure 8b, Table 3). When the pH was adjusted to pH 3.0 (Figure 9b), the aggregation of 537

adsorbed PPM at interface as well as bridging flocculation between the droplets were still 538

evident. To understand whether the reduction of pH had an effect on PPM that were present at 539

the interface, Table 2 shows the adsorption efficiency of the PPM as a function of pH. The 540

PPM had very high degree of adsorption (> 98%) to the droplet surface at all pH (Table 2), 541

with slightly yet significantly higher adsorption at pH 5.0 as compared to those in pH 7.0 or 542

pH 3.0 (p < 0.05). This is also evident visually from the images of the subnantant 543

(Supplementary Figure S4a) after dilution and centrifugation of the emulsions suggesting 544

that the majority of the PPM particles were either adsorbed at the droplet surface or somehow 545

associated with interconnecting the neighbouring droplets in a PPM-PPM network. Overall, 546

E1.0 maintained high stability to coalescence when the pH was adjusted from pH 7.0 down to 547

pH 5.0 or pH 3.0, where the adsorbed PPM on the droplet surface increased aggregation as 548

well as the PPM attached to neighbouring droplets. 549

550

3.3.4 Influence of background electrolyte concentration on behaviour of E1.0 droplets 551

24

With the increase in ionic strength from 1 mM to 10 mM, the droplet size distribution (Figure 552

10a) and corresponding mean droplet diameters (d43, d32) (Table 2), showed no statistically 553

significant differences (p > 0.05). Although the net surface potential of charged droplets 554

became less negative (from - 40 mV to -28 mV), the electrostatic repulsion was still sufficient 555

to inhibit extensive flocculation in E1.0. The emulsions showed shear thinning behaviour 556

irrespective of ionic strengths (Figure 10b). There was no significant difference between 557

viscosities and the n values of these emulsions in the presence of 1 mM and 10 mM NaCl, 558

especially in the region of 0.1-10 s-1 shear rate (p > 0.05) (Table 3). Interestingly, the viscosity 559

of emulsions in presence of 100 mM NaCl was higher than the other emulsions with 560

consequently lower n value and higher K value (p < 0.05) (Table 3). The unaltered adsorption 561

efficiency (Table 2, Supplementary Figure S4b) upon ion treatment (100 mM NaCl) and 562

enhanced viscosity suggested inter-droplet flocculation in E1.0 with 100 mM NaCl, in line with 563

lower net surface charge at the droplet surface (-12 mV) (Table 2). No oiling off or phase 564

separation were observed after rheology measurements, and even after an extended period of 565

storage (Figure 11). Looking closely at confocal micrographs (Figure 11), droplet flocculation 566

can be observed after addition of 100 mM NaCl corroborating with the bulk rheological 567

measurements (Figure 10b). In a previous study, Pickering emulsions stabilized by kafirin 568

nanoparticles showed a reduction in average droplet size on increasing ionic strength from 10 569

mM to 50 mM, which was mainly attributed to enhanced nanoparticle interaction via 570

electrostatic screening effects. On the other hand, de Folter, et al. (2012) suggested that 571

Pickering emulsions stabilized by both positively- and negatively-charged zein particles at very 572

high ionic strength (1 M) aggregated and exhibited an emulsion–gel phase. Comparing our 573

results with these afore-mentioned plant protein-based Pickering emulsions, we hypothesize 574

that a weak gel-like emulsion structure might have been formed which is apparent from the droplet 575

aggregation observed in the confocal images (Figure 11). However, the structure of the O/W 576

25

emulsion might not be as strong as that of a ‘true gel’, but may exhibit a small yield stress, which 577

requires future rheological characterization. 578

579

4. Conclusions 580

Results from our research demonstrate the ability of a new class of plant protein particles i.e. 581

pea protein microgels created using a facile top down approach to stabilize O/W Pickering 582

emulsions with ultrastability against coalescence. To understand the characteristics of these 583

Pickering O/W emulsion droplets as a function of microgel concentration, pH- or salt-584

treatment, the colloidal behaviour of pea protein microgel dispersions in aqueous phase was 585

first investigated at various pH values (pH 2.0 to pH 9.0) or salt concentrations (1 to 250 mM 586

NaCl) was investigated. Aqueous dispersions of pea protein microgels showed highest degree 587

of particle aggregation at pH 5.0 as the activation energy barrier in particle-particle interaction 588

potential was calculated to be extremely low at this pH. Meanwhile, high salt concentrations 589

resulted in charge screening effects in PPM dispersion but the resulting reduction in 590

electrostatic potential did not affect the hydrodynamic diameter of microgels, suggesting that 591

other, possibly steric effects might also be playing a role in the colloidal stability of these 592

particles. Interestingly, when the pea protein microgels were present at the oil-water interface, 593

ultra-stable emulsion droplets were obtained only at microgels with 1.0 wt% protein 594

concentration with all protein subunits i.e. legumin, vicillins and convicillins, being 595

simultaneously present on the microgel-laden interface. The packed layer of PPM particles 596

stabilizing the oil droplets allowed the emulsions to be stable over few months against 597

coalescence. Upon pH reduction to pH 5.0, both intra-droplet and inter-droplet aggregation of 598

PPM occurred resulting in higher adsorption efficiency and higher viscosity, respectively. The 599

emulsions also showed responsiveness to ions, especially at 100 mM NaCl with enhancement 600

26

in viscosity and shear-thinning character. To our knowledge, this is the first comprehensive 601

study that has systematically demonstrated the role of electrostatics in the colloidal stability of 602

the plant-based microgel particles in bulk phase versus particles adsorbed at the surface of the 603

droplets. Findings from this comprehensive study might open door for applicability of these 604

pea protein-based microgels in a range of food products and allied soft-matter applications, 605

where alternative plant-based sustainable Pickering stabilizers are increasingly necessary. 606

Ongoing research is focussing on tuning the surface and bulk properties of pH and ion-607

responsive pea protein microgel-stabilized emulsions, which can help to tailor the in vitro 608

gastrointestinal digestion kinetics of these microgel-stabilized emulsified lipids and allow their 609

usage for controlled release applications in the future. 610

611

Acknowledgements 612

The authors would like to gratefully acknowledge the contributions of Dr Sally Boxal for her 613

technical support in confocal microscopy at the Bio-imaging Facility within the Faculty of 614

Biological Sciences of University of Leeds. 615

References 616

Adal, Sadeghpour, Connell, Rappolt, Ibanoglu, & Sarkar. (2017). Heteroprotein 617 Complex Formation of Bovine Lactoferrin and Pea Protein Isolate: A 618 Multiscale Structural Analysis. Biomacromolecules, 18(2), 625-635. 619

Adebiyi, & Aluko. (2011). Functional properties of protein fractions obtained from 620 commercial yellow field pea (Pisum sativum L.) seed protein isolate. Food 621 Chemistry, 128(4), 902-908. 622

Araiza-Calahorra, & Sarkar. (2019). Pickering emulsion stabilized by protein nanogel 623 particles for delivery of curcumin: Effects of pH and ionic strength on curcumin 624 retention. Food Structure, 21, 100113. 625

Aveyard, Binks, & Clint. (2003). Emulsions stabilised solely by colloidal particles. 626 Advances in Colloid and Interface Science, 100, 503-546. 627

Ben-Harb, Panouillé, Huc-Mathis, Moulin, Saint-Eve, Irlinger, Bonnarme, Michon, & 628 Souchon. (2018). The rheological and microstructural properties of pea, milk, 629

27

mixed pea/milk gels and gelled emulsions designed by thermal, acid, and 630 enzyme treatments. Food Hydrocolloids, 77, 75-84. 631

Binks. (2002). Particles as surfactants—similarities and differences. Current opinion 632 in colloid & interface science, 7(1-2), 21-41. 633

Bora, Brekke, & Powers. (1994). Heat Induced Gelation of Pea (Pisum sativum) 634 Mixed Globulins, Vicilin and Legumin. Journal of Food Science, 59(3), 594-635 596. 636

Chen, Lin, Sun, & Zhao. (2014). Stable and pH-Sensitive Protein Nanogels Made by 637 Self-Assembly of Heat Denatured Soy Protein. Journal of agricultural and food 638 chemistry, 62(39), 9553-9561. 639

Chihi, Mession, Sok, & Saurel. (2016). Heat-Induced Soluble Protein Aggregates 640 from Mixed Pea Globulins and β-Lactoglobulin. Journal of agricultural and 641 food chemistry, 64(13), 2780-2791. 642

Chihi, Nicolas, & Saurel. (2018). Acid gelation of mixed thermal aggregates of pea 643 globulins and β-lactoglobulin (Vol. 85). 644

Cochereau, Nicolai, Chassenieux, & Silva. (2019). Mechanism of the spontaneous 645 formation of plant protein microcapsules in aqueous solution. Colloids and 646 Surfaces A: Physicochemical and Engineering Aspects, 562, 213-219. 647

Cosgrove. (2010). Colloid science: principles, methods and applications (2nd ed.). 648 Chichester, U.K: Wiley. 649

David-Birman, Mackie, & Lesmes. (2013). Impact of dietary fibers on the properties 650 and proteolytic digestibility of lactoferrin nano-particles. Food Hydrocolloids, 651 31(1), 33-41. 652

de Folter, van Ruijven, & Velikov. (2012). Oil-in-water Pickering emulsions stabilized 653 by colloidal particles from the water-insoluble protein zein. Soft Matter, 8(25), 654 6807-6815. 655

Destribats, Rouvet, Gehin-Delval, Schmitt, & Binks. (2014). Emulsions stabilised by 656 whey protein microgel particles: towards food-grade Pickering emulsions. Soft 657 Matter, 10(36), 6941-6954. 658

Dickinson. (2012). Use of nanoparticles and microparticles in the formation and 659 stabilization of food emulsions. Trends in Food Science & Technology, 24(1), 660 4-12. 661

Dickinson. (2013). Stabilising emulsion-based colloidal structures with mixed food 662 ingredients. Journal of the Science of Food and Agriculture, 93(4), 710-721. 663

Du Le, Loveday, Singh, & Sarkar. (2020). Pickering emulsions stabilised by 664 hydrophobically modified cellulose nanocrystals: Responsiveness to pH and 665 ionic strength. Food Hydrocolloids, 99, 105344. 666

Ettelaie, & Lishchuk. (2015). Detachment force of particles from fluid droplets. Soft 667 Matter, 11(21), 4251-4265. 668

Gal, Shani Levi, Tal, & Lesmes. (2013). Emulsions stabilization by lactoferrin nano-669 particles under in vitro digestion conditions (Vol. 33). 670

Hogg, Healy, & Fuerstenau. (1966). Mutual coagulation of colloidal dispersions. 671 Transactions of the Faraday Society, 62, 1638-1651. 672

28

Hu, Yin, Zhu, Qi, Guo, Wu, Tang, & Yang. (2016). Fabrication and characterization 673 of novel Pickering emulsions and Pickering high internal emulsions stabilized 674 by gliadin colloidal particles. Food Hydrocolloids, 61, 300-310. 675

Jayasena, Chih, & Nasar-Abbas. (2010). Functional properties of sweet lupin protein 676 isolated and tested at various pH levels. Research Journal of Agriculture and 677 Biological Sciences, 6(2), 130-137. 678

Jiao, Shi, Wang, & Binks. (2018). High-Internal-Phase Pickering Emulsions 679 Stabilized Solely by Peanut-Protein-Isolate Microgel Particles with Multiple 680 Potential Applications. Angewandte Chemie International Edition, 57(30), 681 9274-9278. 682

Karaca, Low, & Nickerson. (2011). Emulsifying properties of chickpea, faba bean, 683 lentil and pea proteins produced by isoelectric precipitation and salt 684 extraction. Food Research International, 44(9), 2742-2750. 685

Kimura, Fukuda, Zhang, Motoyama, Maruyama, & Utsumi. (2008). Comparison of 686 Physicochemical Properties of 7S and 11S Globulins from Pea, Fava Bean, 687

Cowpea, and French Bean with Those of Soybean French Bean 7S 688

Globulin Exhibits Excellent Properties. Journal of agricultural and food 689 chemistry, 56(21), 10273-10279. 690

Laguna, Picouet, Guàrdia, Renard, & Sarkar. (2017). In vitro gastrointestinal 691 digestion of pea protein isolate as a function of pH, food matrices, 692 autoclaving, high-pressure and re-heat treatments. LWT, 84, 511-519. 693

Liang, & Tang. (2014). Pea protein exhibits a novel Pickering stabilization for oil-in-694 water emulsions at pH 3.0 (Vol. 58). 695

Liu, Ou, & Tang. (2017). Ca 2+ -induced soy protein nanoparticles as pickering 696 stabilizers: Fabrication and characterization. Food Hydrocolloids, 65, 175-186. 697

Liu, & Tang. (2013). Soy protein nanoparticle aggregates as Pickering stabilizers for 698 oil-in-water emulsions. Journal of agricultural and food chemistry, 61(37), 699 8888-8898. 700

Liu, & Tang. (2014). Emulsifying properties of soy protein nanoparticles: influence of 701 the protein concentration and/or emulsification process. J Agric Food Chem, 702 62(12), 2644-2654. 703

Liu, & Tang. (2016). Soy glycinin as food-grade Pickering stabilizers: Part. III. 704 Fabrication of gel-like emulsions and their potential as sustained-release 705 delivery systems for β-carotene. Food Hydrocolloids, 56, 434-444. 706

Makri, & Doxastakis. (2006). Study of emulsions stabilized with Phaseolus vulgaris 707 or Phaseolus coccineus with the addition of Arabic gum, locust bean gum and 708 xanthan gum. Food Hydrocolloids, 20(8), 1141-1152. 709

Meshulam, & Lesmes. (2014). Responsiveness of emulsions stabilized by lactoferrin 710 nano-particles to simulated intestinal conditions. Food & Function, 5(1), 65-73. 711

Mession, Chihi, Sok, & Saurel. (2015). Effect of globular pea proteins fractionation 712 on their heat-induced aggregation and acid cold-set gelation. 46, 233-243. 713

Mession, Roustel, & Saurel. (2017). Interactions in casein micelle – Pea protein 714 system (part I): Heat-induced denaturation and aggregation. Food 715 Hydrocolloids, 67, 229-242. 716

29

Nesterenko, Alric, Silvestre, & Durrieu. (2013). Vegetable proteins in 717 microencapsulation: A review of recent interventions and their effectiveness. 718 Industrial Crops and Products, 42, 469-479. 719

Pan, Tikekar, Wang, Avena-Bustillos, & Nitin. (2015). Effect of barrier properties of 720 zein colloidal particles and oil-in-water emulsions on oxidative stability of 721 encapsulated bioactive compounds. Food Hydrocolloids, 43, 82-90. 722

Peng, Kong, Chen, Zhang, Yang, & Hua. (2016). Effects of heat treatment on the 723 emulsifying properties of pea proteins. Food Hydrocolloids, 52, 301-310. 724

Peng, Xu, Li, & Tang. (2020). Improving the emulsification of soy β-conglycinin by 725 alcohol-induced aggregation. Food Hydrocolloids, 98, 105307. 726

Peng, Xu, Liu, & Tang. (2018). Molecular Mechanism for Improving Emulsification 727 Efficiency of Soy Glycinin by Glycation with Soy Soluble Polysaccharide. J 728 Agric Food Chem, 66(46), 12316-12326. 729

Rayner, Marku, Eriksson, Sjöö, Dejmek, & Wahlgren. (2014). Biomass-based 730 particles for the formulation of Pickering type emulsions in food and topical 731 applications. Colloids and Surfaces A: Physicochemical and Engineering 732 Aspects, 458, 48-62. 733

Sarkar, Ademuyiwa, Stubley, Esa, Goycoolea, Qin, Gonzalez, & Olvera. (2018). 734 Pickering emulsions co-stabilized by composite protein/ polysaccharide 735 particle-particle interfaces: Impact on in vitro gastric stability. Food 736 Hydrocolloids, 84, 282-291. 737

Sarkar, Kamaruddin, Bentley, & Wang. (2016a). Emulsion stabilization by tomato 738 seed protein isolate: Influence of pH, ionic strength and thermal treatment. 739 Food Hydrocolloids, 57, 160-168. 740

Sarkar, Murray, Holmes, Ettelaie, Abdalla, & Yang. (2016b). In vitro digestion of 741 Pickering emulsions stabilized by soft whey protein microgel particles: 742 influence of thermal treatment. Soft Matter, 12(15), 3558-3569. 743

Sarkar, Zhang, Holmes, & Ettelaie. (2019). Colloidal aspects of digestion of Pickering 744 emulsions: Experiments and theoretical models of lipid digestion kinetics. 745 Advances in Colloid and Interface Science, 263, 195-211. 746

Shao, & Tang. (2015). Gel-like pea protein Pickering emulsions at pH3.0 as a 747 potential intestine-targeted and sustained-release delivery system for β-748 carotene (Vol. 79). 749

Tuinier, & De Kruif. (2002). Stability of casein micelles in milk. The Journal of 750 chemical physics, 117(3), 1290-1295. 751

Xiao, Wang, Gonzalez, & Huang. (2016). Kafirin nanoparticles-stabilized Pickering 752 emulsions: Microstructure and rheological behavior. Food Hydrocolloids, 54, 753 30-39. 754

Zhu, Zhang, Lin, & Tang. (2017). Freeze-thaw stability of pickering emulsions 755 stabilized by soy and whey protein particles. Food Hydrocolloids, 69, 173-184. 756

Zhu, Zheng, Liu, Qiu, Lin, & Tang. (2018). Freeze-thaw stability of Pickering 757 emulsions stabilized by soy protein nanoparticles. Influence of ionic strength 758 before or after emulsification. Food Hydrocolloids, 74, 37-45. 759

760